Parental benefits and costs reflect –offspring conflict over the age of fledging among songbirds

Todd M. Jonesa,b,1, Jeffrey D. Brawna, Ian J. Auspreyc, Andrew C. Vitzd, Amanda D. Rodewalde, Douglas W. Raybuckf, Than J. Bovesg, Cameron J. Fissh,i, Darin J. McNeile,i, Scott H. Stolesonj, Jeffery L. Larkini, W. Andrew Coxk, Amy C. Schwarzerk, Noah P. Horsleya, Evalynn M. Trumboa, and Michael P. Warda,b aDepartment of Natural Resources and Environmental Sciences, University of Illinois at Urbana–Champaign, Urbana, IL 61801; bIllinois Natural History Survey, Prairie Research Institute, University of Illinois at Urbana–Champaign, Champaign, IL 61801; cFlorida Museum of Natural History and Department of , University of Florida, Gainesville, FL 32611; dWildlife Section, Massachusetts Division of Fisheries and Wildlife, Westborough, MA 01581; eCornell Lab of Ornithology and Department of Natural Resources, Cornell University, Ithaca, NY 14850; fDepartment of Forestry, Wildlife and Fisheries, University of Tennessee, Knoxville, TN 37996; gDepartment of Biological Sciences, Arkansas State University, Jonesboro, AR 72467; hDepartment of Environmental and Forest Biology, State University of New York College of Environmental Science and Forestry, Syracuse, NY 13210; iDepartment of Biology, Indiana University of Pennsylvania, Indiana, PA 15705; jForestry Sciences Laboratory, United States Department of Agriculture, Forest Service Northern Research Station, Irvine, PA 16329; and kFish and Wildlife Research Institute, Florida Fish and Wildlife Conservation Commission, Gainesville, FL 32601

Edited by Glenn-Peter Saetre, University of Oslo, Oslo, Norway, and accepted by Editorial Board Member Scott V. Edwards October 6, 2020 (received for review May 7, 2020) Parent–offspring conflict has explained a variety of ecological phe- be regarded as a result of interactions between and their nomena across animal taxa, but its role in mediating when song- offspring. Thus, even though offspring may still be dependent on birds fledge remains controversial. Specifically, ecologists have their parents after leaving the nest, the age at which they leave can long debated the influence of songbird parents on the age of depend on the outcome of conflict between parent and offspring fledging: Do parents manipulate offspring into fledging to opti- interests (5). mize their own fitness or do offspring choose when to leave? To In birds, research has shown that the age of fledging (nest provide greater insight into parent–offspring conflict over fledg- – leaving) is primarily associated with nest mortality risk (6 10), but ECOLOGY ing age in songbirds, we compared nesting and postfledging sur- for many taxa, there is compelling evidence that this timing is also vival rates across 18 species from eight studies in the continental mediated by parent–offspring conflict (summarized in ref. 3). One United States. For 12 species (67%), we found that fledging tran- exception to this pattern, however, is in songbirds, where the role sitions offspring from comparatively safe nesting environments to of parent–offspring conflict remains understudied, unclear, and more dangerous postfledging ones, resulting in a postfledging controversial (2). Studies on fledging in songbirds have centered bottleneck. This raises an important question: as past research shows that offspring would benefit—improve postfledging survival—by around a debate over two hypotheses: 1) the parental manipulation staying in the nest longer: Why then do they fledge so early? Our hypothesis (PMH), in which parents and offspring are in conflict findings suggest that parents manipulate offspring into fledging early over the optimal age of fledging and parents initiate fledging by for their own benefit, but at the cost of survival for each individual manipulating offspring (11); and 2) the nestling choice or threshold offspring, reflecting parent–offspring conflict. Early fledging incurred, on average, a 13.6% postfledging survival cost for each individual Significance offspring, but parents benefitted through a 14.0% increase in the likelihood of raising at least one offspring to independence. These Parent–offspring conflict has explained a variety of ecological parental benefits were uneven across species—driven by an interac- phenomena across animal taxa, but its role in mediating when tion between nest mortality risk and brood size—and predicted the songbirds fledge remains controversial. Our analysis of nesting age of fledging among species. Collectively, our results suggest that and postfledging survival rates within 18 songbird species found parent–offspring conflict and associated parental benefits explain var- that offspring commonly leave safer nesting environments for iation in fledging age among songbird species and why postfledging riskier postfledging ones—known as postfledging bottlenecks. bottlenecks occur. This timing of fledging incurs a cost for offspring survival, but benefits by increasing their likelihood of raising at least bottleneck | fledging | parent–offspring conflict | postfledging | songbirds one offspring to independence. Parents therefore appear to manipulate offspring into fledging earlier than expected based onflict between parents and offspring can arise from dis- on the offspring’s ensuing survival prospects. Our results sug- Cagreements over how long periods of parental care should last gest that parent–offspring conflict and associated variation in (sensu, ref. 1). This occurs as extending parental care can enhance parental benefits explain variation in fledging age among offspring fitness, but for adults, the benefits of continuing care are songbird species and why postfledging bottlenecks occur. eventually outweighed by costs to future and survival (2). Consequently, parent–offspring conflict theory proposes that Author contributions: T.M.J. and M.P.W. designed research; T.M.J., I.J.A., A.C.V., D.W.R., C.J.F., D.J.M., W.A.C., A.C.S., N.P.H., E.M.T., and M.P.W. performed research; T.M.J. ana- the optimal length of parental care differs between parents and lyzed data; and T.M.J., J.D.B., I.J.A., A.C.V., A.D.R., D.W.R., T.J.B., C.J.F., D.J.M., S.H.S., J.L.L., their offspring, with offspring preferring care to continue for a W.A.C., A.C.S., N.P.H., E.M.T., and M.P.W. wrote the paper. longer duration than that preferred by adults (1, 3). The length of The authors declare no competing interest. parental care and optimal timing of independence in animals may This article is a PNAS Direct Submission. G.-P.S. is a guest editor invited by the therefore be reflective of complex behavioral interactions between Editorial Board. parents and offspring as they resolve this evolutionary conflict (4). Published under the PNAS license. Though likely strongest over the optimal timing of offspring in- 1To whom correspondence may be addressed. Email: [email protected]. dependence (1), parent–offspring conflict also mediates the timing This article contains supporting information online at https://www.pnas.org/lookup/suppl/ of other juvenile transitions during the period of parental invest- doi:10.1073/pnas.2008955117/-/DCSupplemental. ment. The age at which animals leave their nests, for instance, can First published November 16, 2020. www.pnas.org/cgi/doi/10.1073/pnas.2008955117 PNAS | December 1, 2020 | vol. 117 | no. 48 | 30539–30546 size hypothesis (NCH), in which parents and offspring are not in Pennsylvania, and Texas. First, we estimated daily survival rates conflict and fledging is initiated by other factors only after offspring for juveniles across the nesting and postfledging period and pass a developmental threshold (11, 12). Perpetuating debate be- compared rates immediately before and after fledging. Second, we tween these hypotheses are studies that focus on the proximate used daily survival rates to derive probabilities of mortality for the factors (parental and offspring behaviors) of fledging. For instance, entire brood and compared brood mortality rates before and after past studies have provided evidence in support of the PMH by fledging. Third, we used daily survival rates to estimate cumulative documenting manipulative behaviors of adults, such as luring off- survival for offspring across the postfledging period. Lastly, we spring with food or reducing their provisioning rates (5, 13, 14), but used nesting and postfledging survival rates to derive estimates of other studies have dismissed such observations as anecdotal or parent–offspring costs and benefits (daily survival rates per indi- misinterpretations of behavior (2, 15). Additionally, studies on vidual and per brood) of earlier fledging under scenarios where a these hypotheses have failed to acknowledge an apparent paradox bottleneck does and does not occur. Overall, the goals of our study with respect to fledging. Songbird offspring may be subject to were to use these survival estimates and comparisons to 1) provide postfledging bottlenecks, whereby survival rates immediately fol- evidence to support or refute the NCH and PMH; 2) test for lowing fledging are significantly lower than those in the nest (16), prevalence, severity, and length of postfledging bottlenecks; and 3) with offspring transitioning from a safer nest environment to a more empirically evaluate the influence of parent–offspring conflict on dangerous postfledging one. As experimental and comparative variation in the age at fledging among species by 1) estimating postfledging work shows that songbird offspring benefit in terms of costs and benefits of the age at fledging for parents and offspring development and survival by staying in the nest longer (e.g., refs. 17, and 2) examining associations between nest mortality risk, age at 18), this raises an important question: If offspring are safer and fledging, and costs and benefits associated with parents and off- benefit by staying in the nest, why then, do they fledge so early? spring in this conflict. Consideration of the ultimate factors of fledging (fitness per- spectives of parents and offspring) has the potential to provide Results novel insights into this paradox, the PMH, NCH, and the role of Comparisons of nesting and postfledging survival rates (per- parent–offspring conflict in mediating fledging age. As shown in offspring basis) showed that, although there were cases where a recent paper by Martin et al. (17), offspring survival rates can stage-specific rates were roughly equal (33%), daily mortality be used as proxies for fitness and address the costs and benefits rates of fledglings following fledging (1–5 d) exceeded those of of fledging from the perspective of parents and their offspring. nestlings for most species (67%). Furthermore, probabilities of For individual offspring, fitness is largely dependent on their own brood survival for all species were better outside than within the survival, and selection should favor behaviors that optimize their nest, providing evidence in favor of the parental manipulation chances of survival, such as fledging when nest mortality is hypothesis. Evidence of postfledging bottlenecks was detected roughly equivalent to postfledging mortality risk (11). Parental for 12 of 18 species (Figs. 1 and 2). The severity of bottlenecks fitness, however, is expected to differ from their offspring as their ranged from a 5.7 to 25.3% reduction (mean = 16.2% 1.9 SE) in fitness is influenced by the survival of any offspring within a daily survival upon leaving the nest and lasted for 1 to 5 d brood (17). This is the case as nest predation usually claims an (mean = 2.5 0.3 SE). Bottlenecks occurred across nest types, nest entire brood, while fledglings are dispersed in space and the heights, brood sizes, nestling period lengths, nest mortality risk, brood is usually not entirely predated (reviewed in ref. 19). and habitat type (Fig. 3B and SI Appendix, Table S1). Parents could therefore benefit from manipulating offspring into For offspring, earlier fledging decreased daily survival by an fledging at earlier ages, optimally when the probability of losing average 9.2% (1.4% SE), resulting in an average 13.6% (1.1% the entire brood outside of the nest becomes lower than losing SE) decline in cumulative postfledging survival under a bottle- the entire brood in the nest (17). Benefits for parents in terms of neck scenario compared to if no bottleneck had occurred (see relatively early fledging could come at survival costs to individual Materials and Methods for details on calculations). Under sce- offspring, however, resulting in postfledging bottlenecks. If true, narios where postfledging bottlenecks occurred, earlier fledging this would suggest that parents manipulate offspring to fledge improved the daily probability of at least one fledgling of a brood earlier to increase their own fitness prospects, even at the cost of surviving the immediate days postfledging by an average 6.0% survival to each individual offspring. Furthermore, if such pa- (1.0% SE) relative to if no bottleneck had occurred. Extending rental benefits are uneven with respect to nest mortality risk, these benefits across the postfledging period resulted in an av- then parental benefits may explain the association between nest erage 14.0% (1.8% SE) increase in the likelihood of parents mortality risk and age at fledging among species (6–10). having at least one offspring survive until independence. As nest As such, comparing survival before and after fledging may mortality rates increased, parents derived more fitness benefits provide important evidence for or against the PMH and NCH, (increase in daily brood survival) from offspring fledging at explain the paradox of postfledging bottlenecks, and by proxy, earlier ages (β = 1.165, t = 11.32, P < 0.001; Fig. 3A). Variation empirically evaluate the importance of parent–offspring conflict in nest mortality risk therefore resulted in a gradient of parental in the fledging age of songbirds. Under such comparisons, we benefits following fledging (Fig. 3A), which predicted age at would predict that: 1) if nestlings choose when to leave the nest, fledging among species (β = −50.812, t = −2.61, P = 0.020; for nestling and fledgling survival to be roughly equal and for Fig. 3B). Our post hoc analysis on the effects of brood size, nest nest mortality alone—not parental benefits of fledging—to ex- mortality risk, and their interaction on parental benefits revealed plain age at fledging among species (NCH and no bottleneck); 2) a significant interaction (β = 0.317, t = 5.08, P < 0.001). The if parents manipulate offspring, for nestling survival to be greater benefits of early fledging to adults generally tracked risk of nest than fledgling survival, the probability of losing an entire brood to mortality, but gains were context dependent. Adults with large be lower outside of the nest than in it, and parental benefits—as a broods derived the most benefits by fledging early in risky en- result of nest mortality—to predict the age at fledging across vironments (e.g., low nesting grassland and shrubland species; species (PMH and a bottleneck); fledging ages should benefit Right side of interaction, Fig. 4) and delaying fledging when the parents in terms of brood survival, but come at the cost of post- risk of nest mortality was low (e.g., cavity nesting species; Left fledging survival for each individual offspring. side of the interaction, Fig. 4). Here we use the theory from Trivers’ (1) classic work to provide greater insight into when and why juvenile songbirds leave the Discussion nest. To do so, we compared nestling and fledgling survival rates Our results provide evidence for the parental manipulation hy- for 18 species across eight study locations in Florida, Illinois, Ohio, pothesis and suggest that parents manipulate their offspring into

30540 | www.pnas.org/cgi/doi/10.1073/pnas.2008955117 Jones et al. ECOLOGY

Fig. 1. Per-offspring daily survival rates of juvenile songbirds across the nesting and early postfledging period. Survival rates show clear bottlenecks (>5% drop in survival) during the immediate days after fledging (“postfledging bottleneck”, gray boxes). Dashed lines represent 95% confidence intervals. fledging at younger ages to improve their own fitness prospects, survives until independence. Notably, parental benefits from even at the cost of survival for each of their offspring, reflecting earlier fledging varied among species depending on the inter- parent–offspring conflict. By spatially separating offspring and action between risk of nest mortality and brood size (Fig. 4) and diluting mortality risk to the entire brood earlier, parents can predicted fledging age across species (Fig. 3B). Collectively, our benefit by greatly increasing the odds that at least one offspring results are consistent with findings across other avian taxa that

Jones et al. PNAS | December 1, 2020 | vol. 117 | no. 48 | 30541 Fig. 2. Per-offspring daily survival rates of juvenile songbirds across the nesting and early postfledging period. Survival rates show little to no bottleneck (<5% drop) in survival during the immediate days after fledging. Dashed lines represent 95% confidence intervals.

parent–offspring conflict mediates fledging age among species mortality interact to better align fitness perspectives of parents (summarized in ref. 3) and explain why postfledging bottlenecks and offspring, then parents may gain few to no benefits from occur. manipulating offspring into fledging earlier (e.g., a brood size of Our findings suggest that postfledging bottlenecks occur across 1). Like Martin et al. (17), our findings therefore suggest that the many songbird species (Fig. 1), supporting findings of Naef- paradox between nest mortality risk and fledging age is explained Daenzer and Grüebler (16). Furthermore, our findings show by variation in parental benefits—and presumed manipulation— that bottlenecks vary markedly in their severity and length, can resulting from a key interaction between nest mortality risk and occur regardless of life history characteristics (e.g., nest types, nest brood size. Notably, in some species parents may further dilute heights, brood size, nestling period length, nest mortality risk, etc.) postfledging brood risk by dividing the brood between them (e.g., and habitat preferences, and further corroborate past research refs. 26–28). Though we were unable to assess this for all species demonstrating the importance of the postfledging period for and there may be other benefits to dividing offspring (e.g., pa- first year survival (20), the dynamics of avian populations (21), and rental foraging economics), brood division appears to be the last life history evolution (17, 22). Our results are also consistent with step that parents could take to ensure at least some reproductive past research documenting bottlenecks during juvenile life stages success in a breeding season. across a wide variety of taxa (e.g., refs. 23–25). As mortality risk and clutch size can vary within species, this Past research has suggested that songbird offspring fledge raises an important question: Do parents and their offspring earlier to escape high rates of nest mortality (7, 8); however, such alter their behaviors (i.e., fledging age, manipulative behaviors) behaviors would appear to be maladaptive, given postfledging in response to real-time changes in mortality risk and brood size? bottlenecks found in our study, highlighting an apparent para- Predator playback experiments altering perceived predation risk dox. Furthermore, experimental evidence suggests that offspring (e.g., ref. 29) and experiments manipulating brood size (e.g., ref. would benefit from staying in the nest longer in terms of devel- 30) could be conducted to provide greater insights into this opment and survival (17). As such, a key question remained question. Or alternatively, observational studies could be con- unanswered: If offspring are safer and would benefit more by ducted or compared to investigate differences in the magnitude remaining in the nest, why then, do they fledge so early? Our and duration of bottlenecks within species. For instance, while analysis shows that earlier fledging of offspring accrues greater we found no bottleneck in our ovenbird population, low post- fitness benefits to parents under higher risk of nest mortality fledging survival rates found in King et al. (31) and Jenkins et al. (Fig. 3A), whereas benefits for parents are negatively correlated (32) suggest that postfledging bottlenecks likely occur elsewhere, with offspring fledging age (Fig. 3B). Parental benefits are not or vary annually, in the species. There might also be unique mediated in response to mortality risk alone, however, as spa- opportunities to understand the relationship between nest mor- tially separating more offspring can further dilute mortality risk tality risk and fledging age. For example, Bosque and Bosque (7) to the entire brood after fledging. Thus, if brood size and found that mainland species introduced to safer islands had

30542 | www.pnas.org/cgi/doi/10.1073/pnas.2008955117 Jones et al. (Molothrus ater) provide evidence in favor of the parental ma- nipulation hypothesis. Utilizing a rare breeding strategy known as brood parasitism (placing eggs into other species nests), cowbirds are known to parasitize over 200 host species (33), and as such, their offspring are not evolved for a specific host’s nesting ecology. Thus, under the nestling choice hypothesis, we would expect cowbirds to leave their nest at the same age and stage of devel- opment regardless of the host species, yet cowbirds are known to fledge across a wide range of ages and stages of development (8 to 13 d) (34) SI Appendix,Fig.S1). Additionally, cowbirds left their nests earlier than expected based on evolutionary theory (Fig. 1), and results from a concurrent study suggest the age that cowbirds fledge is largely influenced by the age at which the host young fledge their nests (SI Appendix,Fig.S1). As such, the behavior of cowbird offspring suggests that, like their host siblings, their age of fledging may be mediated by parent–offspring conflict. Parent–offspring conflict may indeed drive postfledging bot- tlenecks and fledging age among songbirds; however, there are several alternatives and caveats that may explain our findings. First, regardless of what age they leave their nest, there is likely an inherent cost in survival for offspring to fledge. All offspring of nesting animals are inexperienced upon leaving the nest, and such inexperience may manifest itself in terms of higher mor- tality while offspring learn to forage for food while avoiding predation. Inexperience appears unlikely to cause bottlenecks alone, however, based on our findings and those of Martin et al. (17). Experiments prolonging the nestling period have shown that offspring can benefit in terms of development and survival ECOLOGY by staying longer in the nest (17). Furthermore, we found that roughly one in three species did not exhibit a bottleneck, pro- viding examples of where inexperience alone was not enough to cause a significant decline in survival. Second, songbird offspring could be timing their fledging to benefit their siblings and im- Fig. 3. Songbird offspring often fledge earlier than expected, when their prove their inclusive fitness (i.e., kin selection) (35, 36). Though risk of mortality is higher outside of the nest than in it. This earlier fledging more study is needed, altruistic acts by nestlings appear unlikely can benefit parents, however, as spatially separating offspring can increase for species experiencing high nest mortality rates and given daily brood survival compared to a scenario where offspring remain in the known rates of promiscuity and extra pair young among songbird nest. Therefore, differentials in nesting (i.e., broods stay in the nest) versus species (37). Finally, a key assumption of our research is that postfledging (i.e., broods leave the nest) daily brood survival reflects fitness nest survival remains constant if offspring would remain in the benefits parents accrue as a result of earlier fledging by offspring. As earlier fledging benefits parents while incurring a cost to individual offspring— nest for the duration of the bottleneck period. While data from implying conflict between the generations—differentials in brood survival our research suggest this is likely the case (constant models best should also reflect parent–offspring conflict over fledging age. Across spe- describe nestling survival for most species), this assumption re- cies (n = 17), differentials in daily brood survival show that (A) earlier mains poorly tested and warrants further inquiry. fledging of offspring accrues greater fitness benefits to parents under Our results and findings from recent studies (17) provide higher risk of nest mortality, and (B) differences in parental benefits predict greater insight into the ultimate factors driving fledging, but the age of fledging among species. Combined, these associations suggest that questions about the proximate factors remain: Are parents parent–offspring conflict mediates variation in the age of fledging among species. Postfledging daily brood survival was averaged for the length of each species’ postfledging bottleneck period (range 1 to 5 d; 1 d for species that did not exhibit a bottleneck). Abbreviations (four letter codes) for study species are as follows: ACFL (Acadian Flycatcher), BRTH (Brown Thrasher), CERW (Ce- rulean Warbler), COGR (Common Grackle), COYE (Common Yellowthroat), DICK (Dickcissel), EABL (Eastern Bluebird), FISP (Field Sparrow), GCWA (Golden-Cheeked Warbler), GWWA (Golden-Winged Warbler), GRCA (Gray Catbird), INBU (Indigo Bunting), MARW (Marsh Wren), NOCA (Northern Car- dinal), OVEN (Ovenbird), RWBL (Red-Winged Blackbird), and WEWA (Worm-Eating Warbler). Calculations of parental benefits can be found in Dataset S1. similar incubation and nestling periods as their place of origin; this raises the question: How many generations might be needed for selection to act on various behaviors associated with age at fledging? Future research will be needed to determine the extent to which bottlenecks vary across space and time and the conse- quences of such variation on the behavioral ecology of songbirds. – Fig. 4. An interaction between daily mortality risk and brood size predicts While we did not examine parent offspring behaviors and thus benefits to parents as a result of earlier fledging ages of offspring. Fitness cannot confirm whether (or how) parents manipulate their off- benefits for parents reflect reduced probabilities of losing an entire brood spring into fledging, our findings on the brown-headed cowbird due to earlier fledging by offspring.

Jones et al. PNAS | December 1, 2020 | vol. 117 | no. 48 | 30543 manipulating their offspring into fledging, and if so, how? Have force fledged, we were able to recapture them and return them to the nest offspring evolved tactics to resist such manipulation (e.g., ref. after processing. After returning them to the nest, we placed a bag over the 38)? Are fledging ages the result of resolved conflict or active nestlings for 5 to 10 min, which was enough to calm them down and for them manipulation by adults (e.g., ref. 39)? And if some offspring to remain in the nest. As a result, we had occasions where nestlings were force fledged, captured, processed, and then returned to the nest and were initiate fledging, what factors are associated with when they resampled (in the nest) the next day (18). To estimate postfledging survival, we leave? To answer these questions, experimental research will be tracked individuals every 1 to 3 d after they fledged using a handheld receiver needed to disentangle behaviors of adults and their offspring. and antenna. We monitored each individual until they either died, dispersed, For example, cross-fostering different aged broods within species or their radio’s battery failed (20 to 60 d depending on the species and allows recording of parental behavior independent of offspring whether an individual survived). We determined fledglings to have died if we age and determination of fledging age independent of the tracked a signal back to a dead fledgling, a predator, or when we were unable amount of parental investment (5). Such experiments could to find a signal (after multiple days and attempts) for fledglings that were too therefore identify potential behaviors parents and offspring use young to have dispersed (18). to manipulate each other and the cues which trigger such be- Statistical Analyses. havior. This includes identifying more hidden aspects of adult To ensure all variables of interest were comparable among species, we reanalyzed all raw data under a standardized statistical behavior that may have been missed in past studies, such as approach. First, we estimated daily survival rates (DSRs) for nests of each parents deserting broods after a set period of parental care (14) species using the logistic exposure method in program MARK or SAS (51–53). or nest size/structure, where adults build their nests so that off- For 1 of our 18 species, the brown-headed cowbird—a brood parasite that spring are forced to fledge due to overcrowding or deterioration. lays its eggs in other species nests to avoid the energetic demands of raising Similarly, fostering broods of species with shorter nestling pe- young (33)—we only used host nest records for which a cowbird egg or riods into those with much longer periods allows for observation nestling was present and included host species as a random effect in our of offspring behavior independent of parental manipulation. logistic exposure model. We note that we included the brown-headed Thus, whether fostered broods stay or leave at the expected cowbird in our study as cowbird offspring are genetically unrelated to their hosts, providing a unique perspective on parent–offspring conflict and fledging age should provide important evidence as to whether the age of fledging among songbirds. For most species (12 of 18), we also fledging age is a result of voluntary behaviors of offspring or calculated DSRs for the incubation and nestling period. Because nest survival active manipulation by adults. may decline as nestlings age due to increased begging and provisioning by adults [reviewed in Martin and Briskie (54)], where possible, we explicitly Conclusions tested for this possibility but only found evidence of decline in two species. Our study provides greater insight into the factors driving when As these declines did not qualitatively alter our findings, we deferred to our and why songbirds leave their nests as well as the prevalence of standardized approach. postfledging bottlenecks. While in a third of cases offspring may We estimated age-specific DSRs of fledglings using multistate models in have initiated fledging, most species exhibited either improved the program MARK. For each species, we reanalyzed postfledging data parental benefits from offspring leaving the nest or a postfledg- following methods in Jones et al. (47). We first assigned fledgling observa- tions to either an alive or dead state, then fixed absorbing states (dead to ing bottleneck, and therefore support the PMH. Studies have dead, dead to alive) to zero and survival probabilities to 1, and estimated commonly identified mortality risk as a critical factor influencing DSRs using transition probabilities (Ψ) where birds remained in the alive the timing and age at which offspring transition from one life state. Past postfledging research has demonstrated fledging age as the main stage to the next (2, 10, 24, 40–44). While mortality risk still plays predictor of fledgling survival in songbirds (reviewed in refs. 16, 21). Thus, a critical role, our findings are in line with those found in other we incorporated age structures before deriving DSR estimates. For each avian taxa (3), suggesting that for birds and other animals with species, we examined a total of 10 models with a priori hypotheses of age parental care, the age of offspring during key transitions may structure predicting DSRs of fledglings. More specifically, we used seven also be mediated by conflicts between parents and their offspring hypotheses derived from past postfledging studies (same among species), (17). For many songbirds, parent–offspring conflict appears to two models based on the timing of fledgling mortality for each species (unique for each species), and a null model (constant survival rate). We se- mediate earlier fledging in response to higher risks of nest lected our top age models based on Akaike’s information criteria adjusting mortality, resulting in key tradeoffs between nest mortality risk for small samples size (AICc) (55), from which we derived our postfledging and nestling period length observed across songbirds (7, 8, 40). DSRs for each species. If the null model was within 2 delta AICc of our top These tradeoffs, in turn, appear to result in a series of events model, we presumed the null model was competitive and best reflected the driving variation in trait development, pre- to postfledging car- age structure of the species. If the null was greater than 2 AICc from the top ryover effects, postfledging bottlenecks, and differential post- model, then we selected the top model based on AICc rank. We also took the fledging survival which explain life history variation among product of these daily survival rates to calculate cumulative rates of survival species (17, 18, 22). Ultimately, our research highlights how in- across the postfledging period (up until 28 d postfledging, a point past – which most fledglings survived and have acquired independence from their corporating estimates of fitness and theory on parent offspring parents) (18). Age-mortality structures for species can be found in SI Ap- conflict can aid in improving our knowledge of the behavioral pendix, Table S2. DSRs by nest stage and fledgling age for all species are ecology of birds and other animals. listed in Dataset S1 in the supplementary materials. Materials and Methods To test for the presence of postfledging bottlenecks, we compared DSRs immediately before (our constant nestling DSRs) and after fledging (the first 1 Though data for our study were derived from 18 species in eight different to 5 d postfledging, which varied by species). We considered a bottleneck to locations, methodologies were consistent across studies, with only a few be present if we observed a 5% or more reduction in daily survival following minor differences in field techniques [see SI Appendix, Table S1 for full fledging, and if the lower 95% confidence interval for nest survival did not species list and associated studies (18, 27, 28, 45–50) for more details]. For overlap the upper 95% confidence interval of postfledging survival by more each species, trained field assistants searched for and monitored nests across than 5%. the breeding season by systematic searching and observing adult behaviors. We calculated costs and benefits of fledging age for parents and offspring For cavity nesting species, we either established or monitored previously using daily probabilities of a whole brood being lost and daily rates of established nest boxes throughout the field season. Nests were monitored offspring survival (per-offspring basis) across the postfledging period, re- every 1 to 6 d until they either failed or fledged young, with nests checked spectively. Specifically, we tested for how mortality would differ if offspring more frequently (every 1 or 2 d) as fledging approached. For each brood would have stayed in the nest (nest mortality rates) versus leaving the nest where at least one nestling fledged, nestlings were radio-tagged several days (postfledging mortality rates). As such, benefits for adults were calculated as before fledging or on the actual day of fledging. In cases where nestlings the difference in daily brood survival in and out of the nest (Fig. 3; nesting fledged prior to our visit, we searched the surrounding vegetation to capture versus postfledging brood survival), while costs to offspring were calculated and tag individuals. For nestlings tagged on the day of fledging, we did our as the difference in per-offspring nestling and fledgling survival (Figs. 1 and best to limit the potential effects of force fledging. In cases where nestlings 2). To compare brood survival in and out of the nest, we used daily nest

30544 | www.pnas.org/cgi/doi/10.1073/pnas.2008955117 Jones et al. survival rates as estimates of brood survival in the nest, as when a nest is phylogenetic corrections altered our results. Results from phylogenetically predated the entire brood is usually lost (19). In contrast, broods are spatially corrected models were qualitatively identical to our original results (SI Ap- separated upon leaving the nest, making brood survival dependent on pendix,TableS4). Given that our findings did not change and applying phy- survival probabilities of individual offspring. Thus, to calculate brood sur- logenetic corrections without adequate evaluation of phylogenetic signal can vival out of the nest we first determined the average DSR during the bot- be inappropriate or misleading (58), we deferred to our original analyses tleneck period via the product of bottleneck DSRs raised to the power of 1 when presenting results. over the bottleneck length: e.g., 2-d bottleneck, X = (DSRday0 × DSRday1) ^(1/2). Brood survival was calculated as the probability that at least one Data Availability. Data for this research are available through previously offspring survives, which is 1 minus the probability that all offspring die; published manuscripts, dissertations, and theses (18, 27, 28, 45–50), and the daily mortality rate during the bottleneck (1−DSRbottleneck) raised the Supporting Information. power of the brood size (^broodsize). Thus, we calculated brood survival out = − − of the nest via the following formula: X 1 ((1 DSRbottleneck)^broodsize). ACKNOWLEDGMENTS. We thank the Florida Fish and Wildlife Conservation We then subtracted brood survival in the nest from survival out of the nest Commission (FWC), the FWC Wildlife Legacy Initiative, US Fish and Wildlife to calculate parental benefits of earlier fledging (for the full set of calcula- Service, Kennekuk Cove County Park, Middlefork Audubon Society, The Ohio tions see Dataset S1). Adult benefits for offspring fledging for species State University Terrestrial Wildlife Laboratory, Columbus Parks and Recre- without bottlenecks were also calculated in the same way; by taking the ation, Franklin County Metro Parks, The Nature Conservancy, The US difference between survival of the brood in and out (day 0 of the post- Department of Agriculture (USDA) Forest Service Northern Research Station, fledging period) of the nest. As cowbird parents do not directly care for their Arkansas State University Department of Biological Sciences, Indiana Uni- versity of Pennsylvania, Blooming Grove Hunting and Fishing Club, Easton offspring during this period, we did not include the species in our analyses Anglers Association, Delaware State Forest, and many private landowners regarding brood survival. provided access to field sites. This research was funded through the Florida Because we found a large difference in the benefits for parents across State Wildlife Grant program (F14AF00892 [T-35]), Federal Aid in Wildlife species, we used general linear models (Proc Glm) (51) to examine inter- Restoration from the Illinois Department of Natural Resources and US Fish specific associations between nest mortality risk, age at fledging, and pa- and Wildlife Service, the USDA National Institute of Food and Agriculture, rental benefits to try to determine why differences occur. As more offspring Hatch Project ILLU-875-963, NSF (DEB-0639429 to A.D.R.), Ohio Division of result in increased dilution of mortality risk to the entire brood outside of Wildlife, the Kirtland Bird Club Ohio Avian Project Initiative, Pennsylvania the nest, species with larger broods should gain a greater benefit from Game Commission and the Arkansas Science and Technology Authority, the earlier fledging as mortality risk to nests increases. Thus, we conducted a US Natural Resources Conservation Service through the Conservation Effects Assessment Project (68-7482-12-502; J.L.L.), Sproul State Forest, Cornell Lab post hoc analysis in which we examine the effects of brood size, nest mor- of Ornithology, the Pennsylvania Game Commission, US Army Corps of tality risk, and an interaction between the two parameters on the parental Engineers, Engineer Research and Development Center, and research grants benefits (improvement in brood survival) of earlier fledging (Proc Glm) (51). from the Illinois Ornithological Society, Association of Field Ornithologists, To account for a potential confound of species size, we included species mass Wilson Ornithological Society, American Ornithological Society, The North as a covariate in all our models. The effect of mass was insignificant and did American Bluebird Society, and Inland Bird Banding Association. This work ECOLOGY not qualitatively change the results for any of our models and was subse- would not be possible without the dedicated efforts of R. Chicalo, S. Jeffreys, quently dropped from our analyses (SI Appendix, Table S3). L. Mengak, J. Brisbane, R. Rhodes, J. Selden, J. Hidalgo, H. Innocent, A. Almond, For our comparative models, we attempted to correct for phylogenetic K. Malachowski, C. Enloe, A. DiGiovanni, N. Suckow, K. Winter, A. Riggs, effects using phylogenetic generalized least squares (PGLS) analyses. For each N. Shah, M. Larrieu, M. Miller, and R. Leeson, L. Kearns, E. Ames, L. Rogers, J. Felix, B. Adams, C. Grimm, S. Rose, R. Zajack, M. Kneitel, A. Petersen, J. Fields, model, we obtained a phylogenetic (consensus) tree via methods in Burleigh K. Mandrekar, J. Dugovich, D. Shustack, S. Lehnen, E. Boves, E. Hingle, et al. (56) and used our estimates in a PGLS analysis with the Caper (57) E. Pokrivka, M. Selvin, J. Simon, L. Strong, D. Watts, R. Veasley, J. Smithmyer, package in R v3.5.2 (R Development Core Team) to test for a phylogenetic A. Mengle, L. Boodoo, A. Marmo, B. Eddinger, C. Campbell-Shall, K. Rogers, signal (λ). For both models, our dataset did not provide enough information F. Rodruigez, J. Geisel, T. Barbee, S. Toner, C. Sayers, E. Moser. S. Hale, M. McGee, to derive a reliable estimate of λ. Thus, we specified PGLS models with λ fixed C. Waas, C. Polhemus, G. Jukkala, L. Dargis, J. Hahn, K. Collins, T. Mangione, at 1 (maximum signal) and compared results with our original models to see if C.Vigneu,K.Ongman,Z.Emery,andS.Freeman.

1. R. L. Trivers, Parent-offspring conflict. Am. Zool. 14, 249–264 (1974). 17. T. E. Martin, B. Tobalske, M. M. Riordan, S. B. Case, K. P. Dial, Age and performance at 2. L. S. Johnson, A. V. Nguyen, C. L. Connor, Do parents play a role in the timing and fledging are a cause and consequence of juvenile mortality between life stages. Sci. process of fledging by nestling mountain bluebirds? J. Field Ornithol. 88,39–46 Adv. 4, eaar1988 (2018). (2017). 18. T. M. Jones, M. P. Ward, Pre- to post-fledging carryover effects and the adaptive 3. M. C. Mainwaring, The transition from dependence to independence in birds. Behav. significance of variation in wing development for juvenile songbirds. J. Anim. Ecol., Ecol. Sociobiol. 70, 1419–1431 (2016). 10.1111/1365-2656.13285 (2020). 4. H. C. Godfray, Evolutionary theory of parent-offspring conflict. Nature 376, 133–138 19. T. E. Martin, A conceptual framework for clutch-size evolution in songbirds. Am. Nat. – (1995). 183, 313 324 (2014). 5. J. Nilsson, M. Svensson, Fledging in altricial birds: Parental manipulation or sibling 20. F. M. Drummond, K. A. Parker, T. G. Lovegrove, D. P. Armstrong, Distinguishing ef- competition? Anim. Behav. 46, 379–386 (1993). fects of juvenile mortality and dispersal on recruitment. J. Wildl. Manage. 83, – 6. R. E. Ricklefs, “Sibling competition, hatching asynchrony, incubation period, and 1744 1752 (2019). lifespan in altricial birds” in Current Ornithology, D. M. Power, Ed. (Springer, 1993), 21. W. A. Cox, F. R. Thompson, III, A. S. Cox, J. Faaborg, Post-fledging survival in passerine birds and the value of post-fledging studies to conservation. J. Wildl. Manage. 78, pp. 199–276. – 7. C. Bosque, M. T. Bosque, Nest predation as a selective factor in the evolution of de- 183 193 (2014). 22. T. E. Martin, LIFE HISTORY. Age-related mortality explains life history strategies of velopmental rates in altricial birds. Am. Nat. 145, 234–260 (1995). tropical and temperate songbirds. Science 349, 966–970 (2015). 8. T. E. Martin, Avian life history evolution in relation to nest sites, nest predation, and 23. M. W. Beck, Size-specific shelter limitation in stone crabs: A test of the demographic food. Ecol. Monogr. 65, 101–127 (1995). bottleneck hypothesis. Ecology 76, 968–980 (1995). 9. V. Remes, T. E. Martin, Environmental influences on the evolution of growth and 24. J. D. Arendt, D. N. Reznick, Evolution of juvenile growth rates in female guppies developmental rates in passerines. Evolution 56, 2505–2518 (2002). (Poecilia reticulata): Predator regime or resource level? Proc. Biol. Sci. 272, 333–337 10. D. A. Roff, V. Remes, T. E. Martin, The evolution of fledging age in songbirds. J. Evol. (2005). Biol. 18, 1425–1433 (2005). 25. L. L. Smith, D. A. Steen, L. M. Conner, J. C. Rutledge, Effects of predator exclusion on 11. L. S. Johnson, R. L. Rauch, S. N. Dellone, The process and causes of fledging in a cavity- nest and hatchling survival in the gopher tortoise. J. Wildl. Manage. 77, 352–358 – nesting passerine bird, the House Wren (Troglodytes aedon). Ethology 110, 693 705 (2013). (2004). 26. C. J. Fiss, “Multiscale habitat selection and movement of fledgling golden-winged 12. C. A. Ribic et al., Diel fledging patterns among grassland passerines: Relative impacts warblers (Vermivora chrysoptera) in two managed mixed-oak forest communities – of energetics and predation risk. Auk 135, 1100 1112 (2018). of Northern Pennsylvania,” PhD thesis, Indiana University of Pennsylvania, Indiana, 13. T. Michaud, M. Leonard, The role of development, parental behavior, and nestmate PA (2018). competition in fledging of nestling tree swallows. Auk 117, 996–1002 (2000). 27. E. M. Trumbo, “Post-fledging ecology of an endangered species: The golden-cheeked 14. T. Grim, Experimental evidence for chick discrimination without recognition in a warbler,” PhD thesis, University of Illinois at Urbana-Champaign, Urbana, IL (2019). brood parasite host. Proc. Biol. Sci. 274, 373–381 (2007). 28. D. W. Raybuck, J. L. Larkin, S. H. Stoleson, T. J. Boves, Radio-tracking reveals insight 15. C. A. Ribic, D. J. Rugg, N. Koper, K. Ellison, C. S. Ng, Behavior of adult and young into survival and dynamic habitat selection of fledgling Cerulean Warblers. Condor grassland songbirds at fledging. J. Field Ornithol. 90, 143–453 (2019). 122, duz063 (2020). 16. B. Naef-Daenzer, M. U. Grüebler, Post-fledging survival of altricial birds: Ecological 29. Y. R. Cheng, T. E. Martin, Nest predation risk and growth strategies of passerine determinants and adaptation. J. Field Ornithol. 87, 227–250 (2016). species: Grow fast or develop traits to escape risk? Am. Nat. 180, 285–295 (2012).

Jones et al. PNAS | December 1, 2020 | vol. 117 | no. 48 | 30545 30. J. N. Styrsky, J. D. Brawn, S. K. Robinson, Juvenile mortality increases with clutch size 44. S. J. Chiavacci, M. P. Ward, T. J. Benson, Why fledge early in the day? Examining the in a neotropical bird. Ecology 86, 3238–3244 (2005). role of predation risk in explaining fledging behavior. Behav. Ecol. 26, 593–600 31. D. I. King, R. M. Degraaf, M. L. Smith, J. P. Buonaccorsi, Habitat selection and habitat- (2015). specific survival of fledgling ovenbirds (Seiurus aurocapilla). J. Zool. (Lond.) 269, 45. I. J. Ausprey, A. D. Rodewald, Postfledging survivorship and habitat selection across a rural-to-urban landscape gradient. Auk 128, 293–302 (2011). 414–421 (2006). 46. N. P. Horsley, “Reproductive success and survival of the common grackle in central 32. J. M. A. Jenkins, F. R. Thompson, III, J. Faaorg, Contrasting patterns of nest survival illinois.” PhD thesis, University of Illinois at Urbana-Champaign, Urbana, IL (2020). and postfledging survival in ovenbirds and acadian flycatchers in missouri forest 47. T. M. Jones, M. P. Ward, T. J. Benson, J. D. Brawn, Variation in nestling body condition fragments. Condor 118, 583–596 (2016). and wing development predict cause-specific mortality in fledgling dickcissels. “ 33. B. D. Peer, J. W. Rivers, L. Merrill, S. K. Robinson, S. I. Rothstein, The brown-headed J. Avian Biol. 48, 439–447 (2017). cowbird: A model species for testing novel research questions in animal ecology, 48. D. J. McNeil, “Population dynamics of species recovery: Multiscale demography in evolution, and behavior” in Avian Brood Parasitism: Behaviour, Ecology, Evolution restored habitats.” PhD dissertation, Cornell University, Ithaca, NY (2019). and Coevolution, M. Soler, Ed. (Springer International Publishing, 2017), pp. 161–187. 49. W. A. Cox, A. C. Schwarzer, B. Tornwall, Postfledging survival of the worthington’s 34. P. W. Woodward, Behavior ecology of fledgling brown-headed cowbirds and their marsh Wren (Cistothorus palustris griseus). Avian Conserv. Ecol. 14, 19 (2019). hosts. Condor 85, 151–163 (1983). 50. A. C. Vitz, A. D. Rodewald, Influence of condition and habitat use on survival of post- 35. L. A. Freed, Forced fledging: An investigation of the lengthy nestling period of fledging songbirds. Condor 113, 400–411 (2011). ’ tropical house wrens. Natl. Geogr. Res. 4, 395–407 (1988). 51. SAS Institute, SASSTAT User s Guide (SAS Institute, Cary, NC, 1990). 36. E. K. Bowers, S. K. Sakaluk, C. F. Thompson, Sibling cooperation influences the age of 52. G. C. White, K. P. Burnham, Program MARK: Survival estimation from populations of – nest leaving in an altricial bird. Am. Nat. 181, 775–786 (2013). marked animals. Bird Study 46, 120 139 (1999). – 37. S. C. Griffith, I. P. F. Owens, K. A. Thuman, Extra pair paternity in birds: A review of 53. T. L. Shaffer, A unified approach to analyzing nest success. Auk 121, 526 540 (2004). 54. T. E. Martin, J. V. Briskie, Predation on dependent offspring: A review of the conse- interspecific variation and adaptive function. Mol. Ecol. 11, 2195–2212 (2002). quences for mean expression and phenotypic plasticity in avian life history traits. Ann. 38. R. A. Johnstone, Begging signals and parent-offspring conflict: Do parents always N. Y. Acad. Sci. 1168, 201–217 (2009). win? Proc. Biol. Sci. 263, 1677–1681 (1996). 55. K. P. Burnham, D. R. Anderson, Model Selection and Multi-Model Inference (Springer, 39. M. Wu, J. Walser, L. Sun, M. Kolliker, The genetic mechanism of selfishness and al- 2002). truism in parent-offspring coadaptation. Sci. Adv. 6, eaaw0070 (2020). 56. J. G. Burleigh, R. T. Kimball, E. L. Braun, Building the avian tree of life using a large- 40. D. Lack, Ecological Adaptations for Breeding in Birds (Methuen, London, 1986). scale, sparse supermatrix. Mol. Phylogenet. Evol. 84,53–63 (2015). 41. T. J. Case, On the evolution and adaptive significance of postnatal growth rates in the 57. D. Orme, The caper package: Comparative analysis of phylogenetics and evolution in terrestrial vertebrates. Q. Rev. Biol. 53, 243–282 (1978). R (2013). cran.r-project.org/web/packages/caper/vignettes/caper.pdf. Accessed 12 42. S. C. Stearns, The Evolution of Life Histories (Oxford University Press, 1992). March 2020. 43. V. H. W. Rudolf, M. Rodel, Phenotypic plasticity and optimal timing of metamorphosis 58. L. J. Revell, phytools: An R package for phylogenetic comparative biology (and other under uncertain time constraints. Evol. Ecol. 21, 121–142 (2007). things). Methods Ecol. Evol. 3, 217–223 (2012).

30546 | www.pnas.org/cgi/doi/10.1073/pnas.2008955117 Jones et al.